Sensitivity Analysis of Support Vector Regression-Based Incremental Capacity Analysis for Battery State of Health Estimations.

Lithium-ion battery state of health (SOH) estimation has become a critical challenge for battery management systems (BMS) in electrified vehicles and consumer electronics. While support vector regression (SVR)-based incremental capacity analysis (ICA) was established as an effective method for estimating SOH, the published work on this topic has been focused on a single set of charging conditions. This paper proposes an improved SVR-based ICA method that expands the working conditions of prior algorithms by incorporating the temperature and C rate sensitivities of ICA explicitly into SOH estimation models. The proposed algorithm is shown to work for different positive electrode chemistries. Moreover, range insensitivity is established for the proposed SVR-based ICA algorithm, which allows the resulting SOH estimation models to be used in various partial charging ranges encountered in practice. Less than 1% root-mean-square SOH estimation error is achieved when the proposed algorithm is assessed using publicly available datasets.